What Should We Eat? Evidence from Observational Studies
Observational studies provide a wealth of important correlations between diet and disease. There is a clear pattern of dietary habits that is associated with reduced rates of a multitude of common illnesses, including heart attack, cancer, stroke, diabetes, and hypertension. In some cases, interventional studies have proven the benefits of dietary change; in others, there is insufficient evidence to prove causation. Based on the existing evidence, the optimal diet should emphasize fruits and vegetables, nuts, unsaturated oils, whole grains, and fish, while minimizing saturated fats (especially trans fats), sodium, and red meats. Its overall calorie content should be low enough to maintain a healthy weight.

There is extensive epidemiologic evidence linking dietary content with a wide variety of illness. Large differences in the rates of heart disease, stroke, diabetes, and cancer have all been linked to food choices. Unfortunately, most of the data is observational, and in many cases there is little direct evidence of benefit from dietary interventions because of the lack of sufficiently powered clinical trials. Dietary intervention trials that measure surrogate endpoints are much more common than those that show direct changes in morbidity and mortality. Nevertheless, it is possible to make evidence-based recommendations regarding diet. The following paragraphs highlight evidence of some of the links between disease and dietary habits derived from observational studies.